Create vertical search for web pages and products with Solr

We want to set up a web and product search resembling Google for a specific segment of some few hundred websites and shops. There are 4 content types: - Web pages (to be crawled e.g. by Nutch) - Products (supplied as XML, Formats: Google Shopping, Amazon, more XML+CSV formats later) - Classifieds (scraping of 6 websites, splitting up approx. 15 fields like start-/end date, price, title, description) - Pictures (OPTIONAL, may be dropped initially - store URL and alt-tag content only) The content will be in 4 languages: EN, DE, FR, IT (more to follow), content language one field The search index will be stored in structured form with up to 20 fields, slightly varying by content type. Web data will need to identify rich snippets with product, breadcrumb and review data. Product data will be loaded daily. Classifieds need to be scraped at least daily. The search engine UI shall get a UI somewhat close to Google search (AJAX-based), providing 4 search areas: Web, Shopping, Classifieds, Images. Classifieds are formatted like Web pages with a mid-sized (approx. 180 px x 150 px) image. The UI will be prepared for translation. The URLs to crawl and the filenames to import for shopping data will be preset in this version (also: no personalisation, LBS, ...). The service (crawler, search engine and web UI) is to be set up with a cloud service (AWS, Azure,..) to be agreed upon. This is an attempt to have the entire service developed in one job. We may split it up depending on your feedback. Please state your experience with the technology involved.